摘要
提出一种新的人脸识别算法.首先,利用主动外观模型(active appearance model,AAM)提取人脸五官特征点,进而获得人脸区域的全局纹理特征;然后对人脸区域中的若干个局部子块进行加权局部二元模式(local binary pattern,LBP)的特征组合;接着分别对这两类特征进行最近邻法则匹配;最后,采用基于模糊综合的原理对这两大类特征进行数据融合,给出最终识别结果.实验表明该算法的有效性,能够很好地结合人脸图像全局和局部的互补信息,识别效果优于各单一模块的分类性能.
A novel method based on the fusion of active appearance model(AAM) and local binary pattern(LBP) for face recognition was proposed.Firstly,we performed AAM to locate facial feature points,after cropping the texture in the facial region according to the shape mesh,the texture was sent to a NN(nearest-neighbor) classifier for recognition.Secondly,a weighted LBP was used to combine the features with various local facial regions(i.e.eyes,eyebrows,nose and mouth).Then these two features were matched to the most dose rules.Finally,we adopted fuzzy integration to fuse both the global texture and local LBP features and the final result was given.Experiments demonstrated the effectiveness and feasibility of our proposed method,which could combine the complementary information of global and local face images,and recognition was better than the classification performance of a single module.
出处
《湖北大学学报(自然科学版)》
CAS
北大核心
2010年第2期146-150,共5页
Journal of Hubei University:Natural Science
基金
广东高校优秀青年创新人才培育项目(LYM08080)资助
关键词
人脸识别
主动外观模型
局部二元模式
特征融合
face recognition
active appearance model
local binary pattern
fuzzy integration